Transducer arrays?

[tohu]

Hello, I've recently begun looking into building my own side-scan sonar and after having looked at a few sites a solution with an array of simple fishfinder transducers seems most appropriate.

However, I have not seen any site that dwelves into optimal spacing of the transducers.

Since the wave-propagation should be similar to that of ordinary antennas I'd expect that it is a matter of displacing the transducers so that the waves cancel out each other in the directions I don't want them to go.

However, things get a bit tricky since the wave-length of a 200kHz signal in water is somewhere between 0,5 and 1 millimeters (dependign on temperarute and salinity) and the transducers themselves are considerably larger than that.

Am I thinking too much about this or is it possible that two or more transducers in a row can actually create a wider omni-directional beam than a singe one?

Also (if at all possible), would I gain from cutting the transducers into a rectangular shape? The thought being that the material that's being cut away wouldn't have contributed more than noise and/or unwanted narrowing of the beam?

Finally, is it feasible or silly to do simulations and mock-ups by using simple (and dirt cheap) piezo-transducers in air? Side-scanning should work in air too, right? (Not using the same hardware, naturally.)

[GLB]

Hi,
My advice it not to waste your money and time with it. Fishfinder transducers are COMPLETELY inappropriate for a sidescan array. (I know, been there, done that while experimenting in the early nineties) The reason is their vibration modes and shapes. If they could work well, the sidescan manufacturer's would use them rather than have the expense of custom made elements. I also know this because I design & build real SSS ducers.
Regards

[Unregistered]

simulations

try here. http://www.beugungsbild.de/sidescan/sonar_index.html

Rene has posted in the past. Do a search

[tohu]

Heh, GLB, that would have been my first choice if they hadn't been so expensive.
Or do you know of any company that sells a sidescan transducer for less than 250 euros?

[GLB]

Hi,
Yes, I am familiar with the 'thinned array'. The problem with the simulations is that they assume an ideal that doesn't work as shown in actual practice. They sure are nice simulations though. You will be very disappointed with the actual results of what you get. Problem is that the fishfinder transducers do not vibrate uniformly off the shelf and it screws everything up at the wave front. I pointed it out to the guy when he first posted it in the original forum. Their shapes also lead to huge side lobes every few degrees that smear everything together. They do not disappear simply by placement as the computer simulations suggest. Like it or not, you gotta get wet diamond precision milling equipment and start cutting. Or pay to have it done. Very expensive in small quantity either way.

[GLB]

Hi Tohu,
You cannot even buy the bare elements for that price. If I may ask, what kind of a sounder were you going to use it on??

[tohu]

My intention is to make one, and to rely heavily on DSPs rather than a jungle of analog stuff in processing the signals...

Being a "pro", have you experimented with focused transducers?

[GLB]

Hi Tohu,
Focused transducers are not something I have experimented with. More difficult and I see no real benefit in this application.

[Unregistered]

Hello Tohu,

You have a decimal error, the wavelength of 200 000 Hz is about 7 mm in water. That fact should make them easier to line up.

I recommend you to try to make the transducer, but not with the fishfinder x-ducers!!!

What do you guys and gals think about cutting thinner slices of rectangular piezo sheets and putting the slices after each other? Connecting them as a standard "Dan's fishfinder sss" and embedding in epoxy?

They are not THAT expencive, even if you buy one for testing.
http://www.piezo.com/prodsheet1sq5A.html

I once pened up a fishfinder transducer and the way it is made is not rocket science.

Good luck!

- R

Hello, I've recently begun looking into building my own side-scan sonar and after having looked at a few sites a solution with an array of simple fishfinder transducers seems most appropriate.

However, I have not seen any site that dwelves into optimal spacing of the transducers.

Since the wave-propagation should be similar to that of ordinary antennas I'd expect that it is a matter of displacing the transducers so that the waves cancel out each other in the directions I don't want them to go.

However, things get a bit tricky since the wave-length of a 200kHz signal in water is somewhere between 0,5 and 1 millimeters (dependign on temperarute and salinity) and the transducers themselves are considerably larger than that.

Am I thinking too much about this or is it possible that two or more transducers in a row can actually create a wider omni-directional beam than a singe one?

Also (if at all possible), would I gain from cutting the transducers into a rectangular shape? The thought being that the material that's being cut away wouldn't have contributed more than noise and/or unwanted narrowing of the beam?

Finally, is it feasible or silly to do simulations and mock-ups by using simple (and dirt cheap) piezo-transducers in air? Side-scanning should work in air too, right? (Not using the same hardware, naturally.)

[tohu]

Hm, one would think that I'm one of those imperial buffoons for not knowing how many millimeters there are in a meter. (sigh) Thanks, 7mm sounds more feasible...

Thanks for the link, but aren't the piezo elements of sonar transducers thicker than that?
The ultra-cheap alternative would be to buy a bunch of simple piezo-elements, such as those from Maplin, and cut them into rectangular pieces:
http://www.maplin.co.uk/Search.aspx?...YU82D&DOY=18m5

Assuming that the wave-theory applies here too you would always get a diffuse beam with round elements since most of the dimensions wouldn't be any multiple of half a wavelength...

Keeping the brainstorm fueled my guess is that elements that are an odd multiple of half a wavelength wide would "leak" the least to the sides.
And also, by placing the individual elements an even multiple of wavelengths apart would make them cancel out whatever is leaked between them.

Heck, I'll run by the store and buy some cheap piezo-transducers and try it out using a 8kHz signal and my ears to measure the beam...

[Unregistered]

OK! send a recording of your test!

Why do you think it has to be so thick, the piezo itself???

Is it not the mass of the whole system that dictates the resonance frequency? Does anyone know the whole theory?

Regarding the spacing of the piezos I do not understand about placing the piezos apart at a certain multiple of wavelengthths... can you explain?
Ideally they should not be spaced apart at all i guess....
If I understand the interference and shaping of a beam, an ideal transducer would be made in one piece, thin and long

As the piezo vibrates it would send the signal from all points along the transducer simultaneously.

All the waves interfere. If the waves are in phase, they strengthen each other, if they 180 degrees out of phase they cancel each other out.

That would build up a single thin, high beam, at the points that are exactly 90 degrees from the transducer. That is the only place where "all" waves are in phase, all other places they cancel each other out.

- R

Hm, one would think that I'm one of those imperial buffoons for not knowing how many millimeters there are in a meter. (sigh) Thanks, 7mm sounds more feasible...

Thanks for the link, but aren't the piezo elements of sonar transducers thicker than that?
The ultra-cheap alternative would be to buy a bunch of simple piezo-elements, such as those from Maplin, and cut them into rectangular pieces:
http://www.maplin.co.uk/Search.aspx?...YU82D&DOY=18m5

Assuming that the wave-theory applies here too you would always get a diffuse beam with round elements since most of the dimensions wouldn't be any multiple of half a wavelength...

Keeping the brainstorm fueled my guess is that elements that are an odd multiple of half a wavelength wide would "leak" the least to the sides.
And also, by placing the individual elements an even multiple of wavelengths apart would make them cancel out whatever is leaked between them.

Heck, I'll run by the store and buy some cheap piezo-transducers and try it out using a 8kHz signal and my ears to measure the beam...

[Unregistered]

Sorry, my bad...

I was apparently wrong about the layer thickness... please check out this site:

http://www.ndt-ed.org/EducationResou...ransducers.htm

http://www.ndt-ed.org/EducationResou...eristicspt.htm

That means that for a 200 kHz the thickness of the active element should be as thick as half a wavelength and the thickness of the "matching layer" should be 1/4 of the thickness.

Now, the question is: is that half a wavelength in the piezomaterial itself? To calculate that we need to know the speed of sound in the element itself right?

Wavelength = speed / frequency

How can we get to know that?

-R

[tohu]

You are right that the piezo-layer neednt be that thick. From some example I read a 56um think layer resonated at 19.8Mhz, which would take a 2,77mm thick layer to resonate at 400Mhz...

However, the simple siren-transducers that are specified to resonate at 4kHz are hardly thicker than 100um...

[tohu]

Thought I'd fill you in on the progress of my experiment:
I've precision cut/milled (using Dremel-lookalike over kitchen sink) three simple transducers (had some left over from an electronic-drum project) and glued them to an aluminium plate, hooked them up to my computer and played a 14939Hz (wl=11mm) signal through them.

At first it seemed like all I had created was an almost painfully irritating tweeter but after adjusting the output level things got interesting.

Before I broke (oops) the array I was able to count five main beams by simply rotating the array horizontally at an armlength in front of my face, between the beams there's silence. Facing the array and rotating it vertically yielded a single steady beam.

Once the glue has set again I'll try to be a bit more scientific using a microphone to measure relative levels so that I can draw some sort of chart of the beams. Even though my creation would make no mother proud, I'll also consider posting a picture of it... :o

[Unregistered]

Wow

OK, Cool!

Right now you have come farther than most people in this forum in building a sidescan transducer!

I guess if you get 5 different beams, they are interfering nicely!!! Now it is just a matter optimizing!

It could be that the beam pattern is much better in water! Remember that the wavelength in water is almost 5 times longer!!!

Keep up the good experiments! How far apart did you space the ducers?

- R